In using a conventional surveillance system, a plurality of cameras are typically used for monitoring and surveillance of different locations of interest, where the video streams generated by the plurality of cameras are summoned to a human operator for deciding if any abnormalities exists. In some instances, the surveillance system may additionally comprise a computer system performing video processing of the video streams, such that the human operator only needs to be involved if the computer system automatically has decided that an abnormality is present, e.g. a possible alarm condition that needs interaction by the human operator.
The introduction of the above mentioned computer system drastically minimized the cost of labor, allowing the human operator to monitor a further plurality of video streams. However, in maximizing the number of video streams handled by a single operator, the work involved tends to be very tedious, thus possibly reducing the monitoring and observation quality. This is specifically valid for repeating spurious alarms taking place at the same location.
Thus, it would be desirable to improve the monitoring condition for the human operator, including reducing the number of spurious alarms presented to the operator.